Cisco announced at Cisco Live APJC that it is integrating Anthropic’s Claude Mythos into its AI Defense platform for adaptive red teaming of AI applications and agents. The addition enables multi-step adversarial testing that plans realistic attack paths while operating only against customer-registered and authorized targets. Sessions are tied to named operators, time-bound, and fully auditable. Alongside this, Cisco is expanding Splunk Agent Observability with Tokenomics, which connects agent token spend to output quality and ROI, plus faster Eval-G1 evaluation models. Cisco IQ now feeds live vulnerability and exposure data into Splunk environments and ServiceNow workflows to speed remediation. Together the updates aim to create a continuous loop for securing, governing, and measuring AI agents at scale. Adaptive red teaming with Claude Mythos is coming soon for select customers.
Cisco announced at Cisco Live APJC that it is integrating Anthropic’s Claude Mythos into its AI Defense platform for adaptive red teaming of AI applications and agents. The addition enables multi-step adversarial testing that plans realistic attack paths while operating only against customer-registered and authorized targets. Sessions are tied to named operators, time-bound, and fully auditable. Alongside this, Cisco is expanding Splunk Agent Observability with Tokenomics, which connects agent token spend to output quality and ROI, plus faster Eval-G1 evaluation models. Cisco IQ now feeds live vulnerability and exposure data into Splunk environments and ServiceNow workflows to speed remediation. Together the updates aim to create a continuous loop for securing, governing, and measuring AI agents at scale. Adaptive red teaming with Claude Mythos is coming soon for select customers.